OpenAI might launch a $200 ChatGPT Pro tier

If you thought $20 a month was the ceiling for an AI subscription, OpenAI just added another zero.

The company officially launched ChatGPT Pro, a new high-end tier priced at $200 per month (or £200 / €229 in Europe). It costs 10 times more than ChatGPT Plus, making it by far OpenAI’s most expensive individual plan.

Unlimited access and an extra-brainy reasoning mode

The existence of the tier was initially spotted hours early by software engineer Tibor Blaho on X, before OpenAI formally rolled out the announcement. For users upgrading from Plus, OpenAI provides a prorated credit for the remainder of their billing cycle.

So what does $200 actually buy? Unlimited access to OpenAI's smartest tools, for starters. Pro members get unmetered usage of the flagship reasoning model o1, the lightweight o1-mini, GPT-4o, and Advanced Voice Mode.

More importantly, the plan unlocks "o1 pro mode." This version allocates significantly more compute to think longer and harder on complex problems. OpenAI says it delivers far more reliable answers in demanding fields like data science, advanced coding, and case law analysis.

To prove its reliability, OpenAI tested the model under a strict "4/4" benchmark: a question is only marked correct if the model answers it accurately four out of four times. Under this test, o1 pro mode beat both the base o1 and o1-preview on competition math (AIME 2024), competition code (Codeforces), and PhD-level science questions (GPQA Diamond).

Because thinking harder takes real time, OpenAI added a progress bar to the interface. If you switch away to another chat while waiting, ChatGPT will send an in-app notification when the answer is ready.

Interestingly, Blaho also spotted a reference on a leaked page indicating ChatGPT Team users would soon get a "limited preview of GPT-4.5," though OpenAI quickly scrubbed that mention.

OpenAI also announced 10 Pro grants for medical researchers at institutions like Harvard Medical School, Boston Children’s Hospital, and Berkeley Lab, supporting work in rare diseases, cancer immunotherapy, and dementia.

Why it matters

We think ChatGPT Pro signals a clean split in the AI market between casual assistants and specialized professional tooling.

For everyday consumers, this plan makes zero sense. If you just need email drafts or conversational help, the standard $20 Plus plan is more than enough. But for quant researchers, software engineers, and academics tackling multi-hour problems, a model that fact-checks itself and avoids subtle hallucinations is easily worth $200 a month. Saving an afternoon of debugging already covers the bill.

Economically, this is a smart pivot for OpenAI. As reasoning models burn through massive compute, flat-rate $20 tiers were always going to buckle under extreme workloads. With reports from The New York Times suggesting Plus could rise to $44 by 2029, and rumors of ultra-expensive enterprise plans circulating for months, OpenAI chose to keep consumer pricing intact while charging heavy compute users what it actually costs.

The real risk here is access disparity. As OpenAI begins gating its most reliable, research-grade models behind a $2,400-a-year paywall, the capability gap between individual tinkerers and well-funded labs is about to get much wider.